EURURO Vol. 72 No. 5

811 860

with otherwise similar risk factors for progression (prior
recurrence rate, tumour multiplicity, size, stage, and CIS).
One of the advantages of the 1973 and 1999 WHO systems
is the ability to identify more aggressive tumours; dividing
HG disease into G2 and G3 may avoid overtreatment
.
Implementation of the 2004/2016 system has been
demonstrated to cause grade migration, with significantly
more Ta cases graded as HG tumours; the resulting costs of
overtreatment (BCG, re-transurethral resection, etc.) and
associated morbidity are unknown .
3.5.2.3. Papillary urothelial neoplasm of low malignant potential.
PUNLMP is defined as a papillary urothelial tumour that
resembles exophytic urothelial papilloma but shows
increased cellular proliferation exceeding the thickness of
normal urothelium . The introduction of this new
category in the 2004/2016 WHO classification aimed to
avoid labelling these patients with the term ‘‘cancer’’ to
decrease psychosocial and economic burdens The
published incidence of PUNLMP ranges from 12% to 39%,
with recurrence rates between 25% and 60% and stage
progression rates between 2% and 8%, very similar to the LG
carcinomas
Ten studies in this systematic review reported a total of
624 patients with PUNLMP and 1303 with G1 tumours
. Tumour recurrence occurred in
75 patients with PUNLMP and 111 G1 tumours (12% vs 9%).
Tumour progression of PUNLMP, defined as any stage
increase, was reported in eight studies
. Progression was diagnosed in six of 354 PUNLMP
patients and 16 of 704 G1 patients (1.7% vs 2.3%).
Progression to muscle-invasive disease from PUNLMP is
very rare; it was found in one of 93 PUNLMP patients (1.1%)
and eight of 250 G1 patients (3.2%).
Our study supports existing data demonstrating that
progression of PUNLMP to muscle-invasive tumour is rare.
The risk of recurrence and stage increase is comparable in
PUNLMP and G1 patients. Moreover, molecular profiles of
PUNLMP and G1 categories are similar Consequently,
patients diagnosed with PUNLMP should be followed up in
the same manner as patients with noninvasive G1 tumours.
3.5.2.4. T1 category. T1 tumours are rarely classified as LG
. As such, the 2004/2016 system does not allow
differentiation of T1 tumours in subgroups with distinct
prognoses
Distribution of 2004/2016 WHO grade in the subgroup of
T1 patients was reported in three studies included in our
systematic review Of 681 T1 tumours, only
13 were classified as LG (1.9%).
Recurrence and progression are more frequent in G3
than in HG tumours. Dividing HG T1 disease into G2 and G3,
a higher recurrence rate (50% vs 68%) was found in one
study and a higher progression rate (12% vs 28%) was
reported in two studies . On the basis of these
findings, the 1973 system may provide more accurate
prognostic information in pT1 tumours. One solution may
be the creation of new classification for grade, including
elements from both 1973 and 2004/2016 systems, as
suggested by van Rhijn et al
3.5.3. Limitations and strengths of the review
Although this systematic review gives the best evidence we
have so far, the quality of the evidence obtained was low,
based on the absence of well-designed prospective studies
with low RoBs. Heterogeneity in study designs, populations,
treatment, definition of progression, incomplete reporting
of outcome data, and lack of individual patient data limited
the analyses that could be done and made meta-analysis
inappropriate.
The main analysis in this systematic review is based on
the studies for which both 1973 and 2004/2016 classifica-
tions were assessed. This approach has minimised bias and
is the major strength of this review. Regarding the
reproducibility part of the review, one study appeared
to present the overall global agreement and global kappa
statistics, and not the agreement between pairs of
pathologists as was done in the other two studies.
Moreover, only two studies with a total of three pathol-
ogists assessed the intraobserver variability between the
1973 and 2004/2016 WHO classifications.
4. Conclusions
The current three-tiered 1973 and 2004/2016 WHO
classification systems for grade are not optimal. Intra-
and interobserver variability are slightly lower in the 2004/
2016 WHO classification but still too high. We could not
confirm that the 2004/2016 WHO classification outper-
forms the 1973 classification in predicting the risk of
recurrence and progression. Each classification identifies
different risk groups of NMIBC patients. In each category of
the 1973 WHO classification (G1, G2, and G3), the risks of
recurrence and progression are higher than in the corre-
sponding category of the 2004/2016 WHO classification
(PUNLMP, LG, and HG). A significant weakness of the 2004/
2016 classification is that it gives almost no prognostic
information in T1 patients, nearly all of whom are classified
as HG. Prospective international multicentre studies and
individual patient data analyses are needed to better assess
the real prognostic value of the 1973 and 2004/2016 WHO
classifications.
Author contributions: Viktor Soukup and Otakar Cˇ apoun had full access
to all the data in the study and takes responsibility for the integrity of the
data and the accuracy of the data analysis.
Study concept and design: Soukup, Cˇ apoun, Herna´ndez, Babjuk, Burger,
Compe´rat, Lam, MacLennan, van Rhijn, Roupreˆt, Shariat, Sylvester,
Zigeuner.
Acquisition of data: Soukup, Cˇ apoun, Cohen, Herna´ndez, Yuan.
Analysis and interpretation of data: Sylvester, Soukup, Cˇ apoun, Cohen,
Herna´ndez.
Drafting of the manuscript: Soukup, Cˇ apoun, Cohen, Herna´ndez, Sylvester.
E U R O P E A N U R O L O G Y 7 2 ( 2 0 1 7 ) 8 0 1 – 8 1 3 811